655 
and the Flagellata . 
3. We must now return to the third line of evolution. 
The Endosphaerine tendency is exhibited by the Protococcaceae 
and Hydrodictyaceae, but we do not find here clear serial 
arrangement of stages in its development. The family Proto¬ 
coccaceae is an aggregate of diversified forms, and almost all 
the genera of one sub-family—the Characieae—have been 
removed elsewhere (cf. Section IV), while Halosphaera has 
a sub-family to itself. In all, however, we find complete 
absence of vegetative division, and individuals are only 
multiplied by formation of zoospores or of gametes. The 
Endosphaereae are mostly strictly unicellular, but Phyllobium 
shows some coenocytic tendencies. 
The Hydrodictyaceae as revised contain only complex 
forms, which are coenobia of coenocytes, and never exhibit 
septate vegetative division ; but the affinities of this specialized 
family with its neighbours are not yet clear. Coelastrum , 
however, is known to multiply its coenobia by vegetative 
division ; but Senn (’ 99 ), in a paper, the most important parts 
of which will be dealt with subsequently, has shown clearly 
that Coelastrum (and with it Sorastrum ) presumably) must be 
transferred to the Pleurococcaceae, leaving only Pediastrum 
to be grouped with Hydrodictyon . In the family thus re¬ 
stricted the coenobia in both genera arise by the apposition 
of motile zoospores within a mother-cell-wall, while all the 
coenobia of the Tetrasporaceae and Pleurococcaceae are 
formed by vegetative septate division of a mother-cell. Thus 
the contrast of tendency on these two diverging lines of 
evolution which I have been maintaining becomes by this 
rectification still more pronounced. 
With very high probability it is this Endosphaerine tendency 
which, carried further, has given rise to the well-defined type 
of the Siphoneae, consisting of a thallus which, though 
essentially coenocytic, is structurally unicellular and lacks the 
solidity acquired by septate cell-division. The possibilities 
of this type of aggregation are considerable, and it has given 
rise to many very varied forms; but as with the Volvocine 
type, nothing appears to have been evolved from it of higher 
x x 
